A generative 3D spatial organization model: exploring potentials for design education

Abstract

Purpose This study aims to explore the potential of computational design approaches, particularly generative 3D modeling, for supporting spatial organization tasks in architectural education. It examines the perceived potential of tools that can contribute to spatial reasoning and design exploration rather than functioning solely as representational media. Design/methodology/approach The research methodology consisted of two parts. In the first, GeoRoomCrafter (GRC), a generative design model, was developed as a visual script-based plugin within the Rhino-Grasshopper environment, combining C# scripting with parametric logic to support 3D spatial organization. The second part involved two structured user studies (ES1-ES2): ES1 provided formative feedback to guide the development process, while ES2 enabled a systematic evaluation of GRC's pedagogical potential and usability through comparative assessment. Findings Regarding perceived usefulness items, student ratings were consistently higher for GeoRoomCrafter than for Magnetizing Floor Plan Generator (MFPG), particularly showing GRC's ability to facilitate more efficient and spatially coherent designs. However, for items concerning perceived ease of use, GRC received slightly lower ratings, indicating usability issues related to the interface's complexity. Despite these interaction challenges, responses to the behavioral intention items showed a clear preference for GRC, suggesting that students recognized its potential benefits. Originality/value This study introduces a novel generative spatial organization model developed as a visual scripting tool to assist novice learners. Unlike existing tools focused on 2D layouts or top-down generation, GRC employs a bottom-up approach with real-time 3D manipulation and terrain integration. This approach addresses the critical gap between abstract diagrammatic planning and tangible, site-responsive volumetric design.